Progress in Colloid Propulsion
by
Jos6 Mariano L6pez Urdiales
M.B.A., College des Ingenieurs, Paris, France, 2002
Ingeniero Aeronautico, Universidad Politecnica de Madrid, Madrid, Spain, 2000
Submitted to the Department of Aeronautics and Astronautics
in partial fulfillment of the requirements for the degree of
Master of Science in Aeronautics and Astronautics
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
September 2004
@2004 Jos6 Mariano L6pez Urdiales, All rights reserved.
The author hereby grants to MIT permission to reproduce and to
distribute publicly paper and electronic copies of this thesis document
in whole or in part.
. . ... . . . . . . . . . . . . . . . . . . . . .
Author ........... ..................., .
Department of Aeronicfs and Astronautics
July 19, 2004
A
Certified by......
...............
Manuel Martinez Sanchez
Professor
............
Thesis Supervisor
Accepted by
MASSAHUSETTS INSTITUTh
OF TECHNOLOGY
FR
S2005
tU~9A RI ES
.....
.............
.N.
..................
A Jaime Peraire
Professor of Aeronautics and Astronautics
Chair, Committee on Graduate Students
Progress on Colloid Propulsion
by
Jose Mariano L6pez Urdiales
M.B.A., College des Ingenieurs, Paris, France, 2002
Ingeniero AeronAutico, Universidad Politecnica de Madrid, Madrid, Spain, 2000
Submitted to the Department of Aeronautics and Astronautics
on July 19, 2004, in partial fulfillment of the
requirements for the degree of
Master of Science in Aeronautics and Astronautics
Abstract
In the early decades of the Space Age, a great deal of work was put into the development of the Colloid Thruster as an electric propulsion system for spacecraft. In spite
of the effort by the end of the 70s the programs were stopped in the USA and Europe
before any design had gotten to fly in space.
An exhaustive study of the literature has been performed to identify what were
the reasons behind the disappearance of Colloid Thrusters. Apart from programmatic
reasons related to the introduction of the Space Shuttle, some technical reasons where
identified. The technical difficulties had to do with the use of arrays of needles.
Aiming at overcoming these difficulties, an alternative way to construct Colloid
Thrusters has been proposed. Instead of needles, holes in Teflon where used. This
has been tested both numerically and experimentally with positive results.
This development may be useful not only for colloid propulsion but also for other
technologies that require electrospray emission.
Thesis Supervisor: Manuel Martinez Sanchez
Title: Professor
Acknowledgments
I would like to thank my sponsor, Fundaci6 La Caixa for providing me with a two year
Fellowship. I am deeply grateful to my advisor, professor Manuel Martinez Sanchez,
for his accessibility and impressive analytical abilities. His insight into the physics
of propulsion is only matched by his humane character and friendliness. I want to
also thank my fellow graduate students at the Space Propulsion Laboratory - Kay,
Shannon, Jean Marie, Murat, Luis, Peter, Paulo, Jorge, Chris, Justin, Yassir, Noah,
James and Tim - for being so competent and helping me whenever I needed it. To all
my friends around the world that have had to endure my rants about spacecraft and
rockets, I appreciate your patience. I want to thank very much my parents, Maria
Eugenia and Jose Juan, for living all the steps in my career as intensely as their own.
The sustained contact and care of my whole family has made my years abroad easier
and more enjoyable. Above all, I want to thank my fiancee, Caroline Tomis, for her
determination and aptitude to make my life wonderful.
Contents
1
11
Introduction
1.1
Rocket propulsion . . . . . . . . . . . . . .
. . . .11
1.2
Colloid propulsion. . . . . . . . . . . . . .
. . . .11
. . . .
13
. . . . . . . . .
. . . .
14
1.3.1
Starting voltage . . . . . . . . . . .
. . . .
14
1.3.2
Hydraulic impedance . . . . . . . .
. . . .
16
1.2.1
1.3
Electrospray, principle of operation
Theoretical considerations
19
2 Literature review
2.1
Introduction of the concept and first efforts . . . . . . . . . . . . . . .
19
2.1.1
Colloid propulsion systems not using electrospray . . . . . . .
20
2.1.2
Initial US electrospray-based colloid thruster development
. .
21
2.1.3
Initial European electrospray-based colloid thruster development 29
2.1.4
Soviet electrospray-based colloid thruster development
2.1.5
Contemporary colloid thruster development
. . . .
31
. . . . . . . . . .
32
3
The needle-less thruster
37
4
Simulation
41
. . . . . . . . . . . . . . . . . . . . . . . .
41
. . . . . . . . . . . . . . . . . . . .
44
4.3
Sim ulation results . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
44
4.4
Effect of the dielectric constant on the field enhancement . . . . . . .
50
4.1
Statement of the problem
4.2
Simulation method and accuracy
5
5
Experimental verification
53
6
Further work
55
7
6.1
Exposure to the space environment . . . . . . . . . . . . . . . . . . .
55
6.2
Hydraulic impedance . . . . . . . . . . . . . . . . . . . . . . . . . . .
56
59
Conclusions
6
List of Figures
1-1
Functional diagram of an electric rocket
12
1-2
Electrospray off. . . . . . . . . . . . . .
15
1-3
Electrospray starting......
15
1-4 Electrospray on. . . . . . . . .
15
3-1
Array of holes in a solid block
39
4-1
Case I dimensions . . . . . . . . . . .
42
4-2
Case II dimensions
. . . . . . . . . .
43
4-3
Case I: Convergence
. . . . . . . . .
45
4-4
Case I: Electrostatic potential
4-5
Case II: Electrostatic
4-6
Electric field modulus for a needle, maximum electric field of 3.53106
V/m.
4-7
# for a metallic needle
potential 4 for a hole in Teflon
46
. . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
48
Electric field modulus for a teflon hole, maximum electric field of
3.97106 V /m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-8
47
49
Model grid. For this particular run the dielectric material was teflon,
the hole had 0.4 m of radius and was placed 17.25 mm from the extractor plate.
4-9
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
Close up of the electric field modulus near the tip. For this particular
run the dielectric material was teflon, the hole had 0.4 m of radius and
was placed 17.25 mm from the extractor plate. . . . . . . . . . . . . .
51
4-10 Effect of the relative permittivity on the Electric field at the tip. . . .
52
7
5-1
Taylor cone formed on a hole on Teflon . . . . . . . . . . . . . . . . .
6-1
Close up of the electric field near the meniscus.
54
The electric field
intensity that would be attained in the absence of the conductive film
is 1.1 107 V /m . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6-2
57
Top: Cylindrical emitter hole. Bottom: Flow restriction at the outward
end of the cylindrical emitter hole to increase the flow impedance.
8
58
List of Tables
2.1
Performance of a 432 needle prototype in 1974.
9
. . . . . . . . . . . .
28
10
Chapter 1
Introduction
1.1
Rocket propulsion
A rocket is device that contains matter and a means to provide some part of this
matter with kinetic energy and then eject it in a desired direction and in a controlled
rate. The remaining mass of the rocket undergoes a time rate of change of momentum
which constitutes a useful thrust.
An electric rocket is a type of rocket which transforms electric energy into kinetic
energy for the ejected matter. This concept is illustrated in figure 1-1
A good review of the technologies and state of the art in electric propulsion of
spacecraft can be found in reference
1.2
[47].
Colloid propulsion
Colloid propulsion is a type of electric space propulsion . As such its aim is, using
electrical energy, to expel at great speed a jet of matter that is globally electrically
neutral.
In its operation particles with a high electric charge per unit mass are produced.
These highly charged particles are accelerated through an electrostatic potential. This
same phenomenon is called clectro-spray.
A colloid is a fine dispersion of particles. The matter in the jet that is produced
11
Propellant
Reservoir
Thruster
Figure 1-1: Functional diagram of an electric rocket
12
from a colloid thruster is finely dispersed and hence the denomination of "colloid
propulsion".
In most rocket propulsion systems the exhaust resembles a gas and
therefore its behavior, notably its expansion, is different to that of a colloid thruster.
In order to assure the electrical neutrality it is necessary to expel the same quantity
of charge that is being accelerated but of an opposite sign. This can be achieved either
with many jets, some positively charged and others negatively, or by alternating the
charge emission or finally by emitting a single positive jet and using an auxiliary
cathode to emit electrons.
Of the many ways to produce charged particles that have been studied for this
application it is electro-spray that seems to offer the best efficiency and uniformity. [18]
Electro-spray is a physical phenomenon that is used for many applications aside
from electric propulsion. For instance, electro-spray is used to paint cars, or in the
detection, by mass spectroscopy, of molecules in a sample. The idea is to subject a
mass of electrically conductive liquid to a strong electric field. The field will draw
charges to the surface of the liquid. If the field is strong enough, then the self repulsion
due to the surface charge density overcomes the surface tension that holds the liquid
together and, in this way, small charged drops will be ejected from the mass of liquid.
The liquid should contain both negative and positive free charges.
Two types of liquids have been exploited as propellants. On the one hand, solutions of salts, such as salt water that has Na+ and Cl- dissolved in water and, on the
other hand, molten salts. The solutions that are often used are based on formamide
or glycerol. Liquid water, although an excellent solvent, is troublesome when used in
a vacuum environment.
There exist salts that remain liquid at room temperature in a large window of
pressures, they are known as ionic liquids. An example that is often used in colloid
thruster studies is 1-ethyl-3-methylimidazolium tetrafluoroborate (EMI-BF4).
1.2.1
Electrospray, principle of operation
The following diagram shows an scheme of how an electro-spray works. In an axially
symmetric configuration, an orifice full of liquid is in front of an extracting ring.
13
"
In figure 1-2 the circuit is open and the liquid has its positive and negative
charges evenly distributed. There are no external forces applied on the liquid
and therefore the surface tension minimizes the surface of the liquid.
" In figure 1-3 the circuit is closed and a potential of about 1500 V is applied between the extractor ring and the liquid. Negative charges are attracted towards
the liquid surface. Electric forces acting on the charges on the surface deform
its shape into a cone, known as the Taylor cone.
" In figure 1-4 the surface charge density is large enough for negative charges to
leave the bulk of the liquid towards the extractor. The electrospray has started.
1.3
Theoretical considerations
Some concepts that will be used later on this work will be introduced in this section.
An extensive review of the magnitudes and physics involved in colloid thrusters can
be found in reference [12].
1.3.1
Starting voltage
In order to start the emission of charged matter from an electrospray a certain electric
field must be induced on the surface of the liquid.
Using an order of magnitude analysis and assuming the pressure in the liquid is
small, one can equate the electrostatic pull felt on the surface with the surface tension.
The following expression for the field can be thus obtained.
sE3~,
2-/
2
Re
_____-
2~y(1.1)
In equation 1.1 E, is the vacuum permittivity, Eu, is the electric field at the tip,
'
is the surface tension and Rc is a principal radius of curvature of the surface at the
tip.
14
Extractor
High
Voltage
Needle
I
Figure 1-2: Electrospray off.
Extractor 4
0
h
High
T
Voltage
Needle
Figure 1-3: Electrospray starting.
Extractor
0
aIIj~
High
Voltage
Needle
15
Figure 1-4: Electrospray on.
Assuming the shape of the tip constitutes an equipotential hyperboloid of revolution and the extractor is a plane the value of the electric field at the tip can be
related to the voltage between the tip and the plane. Equation 1.2 was derived originally in the study of the field around solid metal tips [16]. This expression does not
take into account the fluid dynamic nature of the phenomenon and it is therefore an
approximation.
VVst art = rg
loIn( dc )
(1.2)
12
In equation 1.2, de is the diameter of the meniscus and D is the distance from the
extractor to the meniscus.
1.3.2
Hydraulic impedance
One of the most important aspects of colloid thruster development is how to integrate
a large number of emitters so that the total thrust approaches the values of interest
for application of the technology. Arrays of emitters are needed to achieve this.
There are two ways an array of emitters can be fed with propellant. On the one
hand each emitter could have its own propellant reservoir. This is an approach that
has not been yet explored and holds promise because of the advances in microfluidics
and microfabrication. On the other hand emitters may share a common reservoir and
be fed from a manifold. This second approach is the most common one and the one
that will be followed in this study.
It is useful to introduce the concept of hydraulic impedance, Zh, across a duct.
It is defined in equation 1.3 as the pressure difference across the duct (AP) between
the mass flow flowing through it (r1).
Zh Zh AP
m
(1.3)
It is generally accepted that in order to obtain stable emission from each emitter
the hydraulic impedance from the common manifold to the tip of the emitter has to
be large for each emitter. Otherwise only the emitters with low hydraulic impedance
16
would emit. This implies that designs will typically have large ratios of the length of
the channel divided by its cross sectional dimension.
17
18
Chapter 2
Literature review
The literature concerning colloid thruster development has been reviewed. It was
found that while, a great deal of work was devoted to this subject in the 60s and 70s,
there is a blackout in the 80s and early 90s. In addition, it has been noted that the
work done from the mid-90s on is not a continuous development of what was left in
the late 70s. This review has been done with the aim in mind of finding out why
the development of this technology suddenly stopped. Shedding light on the reasons
that lead to that interruption is a useful step in the process of directing innovation
to improve the system.
2.1
Introduction of the concept and first efforts
The concept of electrostatic acceleration of colloids started to be seen as a viable and
attractive means of propulsion in the early sixties.
[64]
Mission designers saw this
technology, in synergy with nuclear reactors, as an enabling technology for human
exploration of the Solar System.[341
Different concepts to create the charged colloid were presented and studied. They
produced a stream of charged particles by different means. These methods will be
exposed in section 2.1.1.
19
2.1.1
Colloid propulsion systems not using electrospray
A great deal of the effort was devoted on the generation of the charged colloid itself.
And it was not taken for granted that it had to be done using electrospray.
Some of the efforts to produce the charged colloid involved using small preformed
solid carbon particles. In 1965 the results of an experiment were presented. Carbon
particles were injected inside an Argon cold cathode ion source. The system applied
an anode voltage of 1350 V and -1000 V at the cathode. The thrust was deemed too
small to be practical.[23]
A different way to charge the carbon particles was tested with a new design. It
used a hot filament to bombard with electrons a metal anode. This produced soft
X-rays that by X-ray photoelectric effect and Compton scattering charged positively
the particles. Two ring electrodes axially mounted provided the axial electric field
that accelerated the charged particles. The specific charge was measured and a distribution between 1000 C/Kg and 10000 C/Kg was obtained. It was seen as a potential
attractive alternative to cesium ion engines.[24].
At NASA Lewis the charging of preformed submicron carbon powder was studied.
The particles were charged by contact between two metallic electrode plates at high
voltage. Particles oscillated when they are charged between both electrodes. One of
the electrodes had a hole through which some charged material eventually escaped.
The powder was of about 1.5 pm average size. The specific charge was detected with
a quadrupole mass filter. This device consists basically in four cylindrical rods in
the corners of a square. Alternating current is applied to the rods to destabilize all
particles trajectories except those with a particular q/m which is collected at the end.
The specific charge measured was much lower than the one that would correspond
to a sphere charging in contact with a plane. The reason for the poor charging is
presumably that the once the particles charged a bit they left the contact with the
electrode due to electrostatic repulsion, before acquiring that maximum charge.[21]
At Stanford, some efforts were dedicated to improve the mechanism of charging
of the solid particles with hot cathodes. A solenoid was added to trap the electrons
20
and enhance the charge transfer to the particles. In this new design, particles became
charged because of secondary electron emission. Aluminum Oxide particles of 1 Pm
of diameter where used. The specific charge obtained was in the range 100-200 C/kg.
The device operated with voltages of 700V and -1000V. [281.
Thiokol, the leader in Solid Rocket Propulsion in the United States of America
(currently ATK) expressed interest in the concept of colloid propulsion.
Thiokol
proposed to generate a charged colloid beam by chemical reactions. A diffusion flame
of Mg vapor with 02 produced nanoparticles of MgO. Then a hot filament was used
to charge them. This was enclosed in a tube of 3 inches of diameter. The nominal
massflow of the thruster was 0.6 g/min .[13]
At NASA, experiments were done on the production of liquid particles from condensation of a homogeneous vapor. Vaporized propellant was expanded and rapidly
cooled in a supersonic nozzle and thus homogeneously condensed into particles of 5-10
nm. The problem that approach faced was that the particle formation efficiency was
too low.[21]
All these efforts to generate highly charged and monodisperse powders failed to
reach the ranges of q/rn that would make them attractive for colloid propulsion.
However it should be noted that in recent years significant progress has been made on
the capabilities to produce very small monodisperse solid particles. The growing field
of nanotechnology may provide ways to overcome the limitations faced by researchers
in the 60s and render the concept attractive again.
2.1.2
Initial US electrospray-based colloid thruster development
The idea of building a colloid thruster based on electrospray of a liquid was already
discussed in the early 60s. Three patents protecting three particular embodiments
of the technology, were issued in 1964. These works only provided general advice
on the properties that the liquid propellant should have, such as low vapor pressure. It was yet unclear then what liquid and with what dopants would be most
21
appropriate. [61] [60] [20]
The United States Air Force started a systematic effort to identify a propellant
that would work well in the negative mode, that is when the jet is negatively charged.
A concern was to avoid negative emission of electrons since that would greatly decrease the efficiency. They identified a set of electrochemical criteria that the metallic
electrode must have with respect to the ions in solution in the propellant. A series
of time of flight experiments was carried out. One propellant combination that performed well was ZnCl 2 dissolved in glycerol, resulting in a specific charge of over 3000
C/kg. [72].
Also interested in the negative mode operation, TRW went on to do a parametric study using a singel needle as an emitter. They looked for working fluids that
performed well on the negative mode. An extensive campaign of propellant testing
on negative mode was performed. The propellants were all glycerol based and the
dopants included H2 SO 4 , MgCl 2 FeCl 3 and SnCl 4 . The most promising results came
from solutions of glycerol doped with Stannic Chloride (SnCl 4 ) and glycerol doped
with Ferric Chloride FeCl 3 . Apparently the chloride plays a role in avoiding the formation of bubbles in the negative mode. SnCl 4 was chosen to be used for the negative
mode in a bipolar design. Using time of flight, a specific charge of up to 3000 C/kg
was measured. The typical massflow was 1.3 10- 9 kg/s [30]. On all negative modes
erosion was seen after more than 200 hours. [29] TRW presented a working prototype
of a 16 needle bipolar module. In order to test the new colloid thrusters, TRV built
a vacuum thrust test-stand able to measure in the range 1-8 plb.[76] The device had
an overall specific impulse of 800 s at an efficiency higher than 70%. The neutrality
was measured as the voltage on the collector. This voltage remained always in a
range of ±100 V. The thruster used different propellants for each mode of operation,
for the positive one it used 19.3 % by weight solution of Nal in glycerol, whereas for
the negative mode it used SnCl 4 . The propellant feeding mechanism consisted on a
spring loaded tank. The thruster was capable of electrostatic thrust vectoring and
the transversal thrust was measured. The engine compared favorably to cesium contact ion engines and was proposed for North-South station keeping of geostationary
22
satellites. [7]
In a joint effort of the Air Force Aero Propulsion Laboratory in the WrightPatterson Air Force Base in Ohio and TRW, a bipolar thruster prototype was built.
It sprayed simultaneously with eight negative emitters and six positive ones. Time of
flight techniques were used to measure a thrust of 51 plb at an specific impulse of 760
s. Nal/glycerol was used for positive emission while SnCl 4 /glycerol was chosen as the
propellant for the negative emission. This negative mode propellant was developed to
mitigate the bubble formation problems of NaI/Glycerol when used in the negative
mode. The gas in the bubbles was identified to be Hydrogen. Stannic Chloride was
chosen because the stannic ion reduced itself to stannous thus oxidizing the hydrogen
which this way stayed in solution. Among other compounds that had this property
SnCl4 was preferred for the higher conductivity solutions that could be prepared with
it. Another problem that was encountered was arcing. Positive changes in the design
were: to reduce capacity of the needle-emitter by reducing the mass of the extractor,
and to place a 105 Q resistor between each emitter and the high voltage circuit. Three
life tests of about 50 hours were performed. To check the neutrality of the beam, the
current impinged on a floating collector that stayed always at ± 100V. [1]
The Air Force of the United States of America along with Electro-Optical systems,
a subsidiary of Xerox Corporation, worked on another exploratory program of colloid
propulsion. The study mapped the specific charge distribution efficiency with average specific charge for a large group of propellants. A general trend followed by most
propellants tested was that increasing the average specific charge of the beam had a
detrimental effect on the charge distribution efficiency. The best propellant identified was tetraethylammonium chloride oleic acid (TEAC). TEAC performed equally
well for negative and positive voltages, it also exhibited good feed control and reproducibility. An attempt at a bipolar thruster was made. Tapered needles with a sharp
rim of 0.001 inches were chosen for the injector geometry. The needles were made
with platinum to prevent corrosion. The needles' capillaries were stuffed with a wire
to increase the flow impedance. The spacing between similar polarity emitters was
0.2 inches and 0.141 inches between opposite polarities. The propellants used were
23
20 g of NaI in 100 ml of glycerol for positive operation and 2 ml H2 SO 4 in 100 ml
glycerol for negative. The emitter section was designed to operate at a mean negative
voltage to shield it against electrons. The voltages of operation were 4.4 kV and -5.8
kV. The average currents in the experiments were 150 pA plus and -154 pA negative.
The final output was a bipolar thruster of 70 plb of thrust what worked for over 50
hours. Using time of flight experiments the power over thrust ratio was measured to
be 20 W/mlb.[50] In parallel, at the United States Air Force Base of Wright-Paterson
work on a 80 1lb bipolar thruster was initiated. The typical specific charges achieved
were 1.3 105 C/kg in the positive mode and 1.3 104 C/kg in the negative mode. The
voltages used were also +4.4 kV and -5.8 kV. It was claimed that the system yielded
better efficiency than ion engines up to an specific impulse of 3000 s. The inherent
scalability of the device promised a large array of potential uses. Another advantage that was pointed out was that it could operate in pulsed mode without power
penalties. The main problems that were encountered were erosion of the emitters
due to arcing as well as deposition on the negative emitters and erosion on positive
ones. The system used 73 needles as emitters. The positive mode propellant was a
solution of 20g of Nal in 100ml of glycerol, whereas the negative mode propellant was
2ml of H2 SO 4 dissolved in 100 ml of glycerol. [15] As part of this effort, models and
analytical derivations of performance curves for these thrusters were developed.[53]
From experimental observations it was understood that the satisfactory operation of the thruster greatly depended on the needle electrochemical properties and
geometry. NASA at Goddard started a project aiming for a better understanding of
what should drive the needle design. In an extensive test program, forty-eight single
needles with various materials and geometries were tested. Pointing out the analogy
of electrospray and an electrochemical process in an electrolytic cell, needles where
tested as electrodes on an electrolytic cell. To accelerate the corrosion process the
current applied was of several mA instead of the pA encountered in electrospray operation. Experiments on a electrochemical cells were performed at 20 V DC. The fluid
in the cell was a typical glycerol Nal colloid thruster propellant. The experiments
showed that heavy deposits of iodine appeared on gold plated stainless steel needles
24
while, on the platinum plated ones, the iodine washed off easily. Recognizing the difficulty of maintaining aligned an array of hundreds of needles permanently attached
to a common feed system, all equidistant from each other and each surrounded by
an extractor, considerable thought was given to finding an alternative and simpler
geometry. An annular slit thruster was then proposed, built and tested. 'The propellant was made to flow out of the gap between two concentric surfaces. It was tested
only in the positive mode. The propellant was 25 g of Nal in 117 ml of glycerol. The
propellant was fed at a pressure of 55 mmHg. The total voltage was 20 kV and it
extracted a current of 175 ptA. The specific charge achieved was 2852 C/Kg, yielding
a thrust 6.2 plb at an specific impulse of 1085 s. [68]
The design of the annular thruster was further improved in NASA Goddard. The
annular gap between the concentric cylinders was brought down to 0.001 inch or
less. [74]
A series of tests on a 36 needle thruster were performed. The device was capable of
thrust vectoring. Normal conditions of operation where the following: total voltage
of 9.35 kV, current of 240 pA, feed pressure of 1.1 inches of Hg, massflow of 12
pgram/s. In that conditions an specific impulse of 1650 s at an efficiency of 70%
where attained. The performance was measured both by thrust stand measurements
and time of flight. The agreement between both methods was up to a 10%. [31]
Aiming for high thrust controllability, TRW presented a three needle thruster. It
consumed 5 W of power to provide 8 plb of thrust at 990 s of specific impulse. It had
a very smooth steady state and transients. The total system weight was 7 lb. The
device operated for 1750 hours with a current of 36 pA. It was noticed that operating
at higher voltages a more collimated beam was produced.
A study of the beam divergence was performed. It was found out that the specific
charge emitted from the needles was anisotropic. Therefore the time of flight values
measured using one single collector provide only the average value of the charge
distribution but not the value at each angle of divergence. [19]
TRW also explored AC and pulsed operation of colloid thrusters. It was hypothesized that the hydrogen bubble generation problem present when operating in the
25
negative mode is neutralized by the free iodine produced by the positive half of the
cycle. The minimum pulse periods measured where of 50 ms. It was observed that to
operate in pulsed mode, it is better to reduce the extraction voltage by about 25 %,
just below the voltage at which the emission stops, instead of turning it off. Interesting observations of the AC transient jet dynamics where made: the jet collapses in 1
ms however to completely reform the jet the time needed was about 3 ms. Results
were taken using both square and sinusoidal waves. In spite of having an inferior
duty cycle, sinusoidal AC was preferred over square waves because of simpler and
less heavy electronics. The results indicated a lower specific impulse for the negative
half period. For instance, at 50 Hz the specific impulse in the negative mode was
about 500 s and in the positive about 700 s. Using stroboscopic imaging it was seen
that, if switching was done at less than 100 Hz, the jets appeared and disappeared
during switching. For higher frequencies the jets remained stationary through the
cycle. Different frequencies up to 50 kHz were studied. The idea behind going to
very high frequencies was that, at some very high frequency, the thruster would not
be space charge limited. The mechanism proposed was that the ions and particles
would make a quasineutral plasma and, as such, it would not build up potentials high
enough to spread the beam or stall it. [62]
In another series of tests, a single capillary needle in AC was used. An audio
oscillator fed a sine wave signal to a 200 W amplifier which was connected to a high
voltage transformer with 4 kV rms. Again, while recognizing that square wave is
more attractive because of its fuller duty cycle, sinusoidal waves are preferred because
the simplicity of the analog electronics required outweighs the loss in efficiency. In
the study different propellants were used, including Nal with glycerol, tetraethyl
ammonium chloride and H2 SO 4 with glycerol. The range of frequencies studied was
from 100 Hz to 2.7 kHz. It was noticed that in the continuous negative dc mode Nal
performed "hashily".
On the other hand at AC the operation was very smooth at
both half cycles. Both the positive and negative beams have particles of an specific
charge of the order of 10000 C/kg at 1000 s of specific impulse. Using time of flight
it was determined that the efficiency was about 40%, which may seem low but since
26
no neutralizer would be needed it is still considered attractive for thrusts below 100
plb. [11]
The United States Air Force wanted to increase the thrust per emitter and decided
to explore two new configurations of emitters. One was a tubular emitter that has
a capillary for mass flow control ending on a tip that has a rim of relatively large
diameter was constructed. A four emitter prototype was tested a 100 hour resulting in
75 %efficiency, a current of 265 pA and 123 plb of thrust. The other new configuration
consisted in annular feeds between two concentric cylinders. A three emitter prototype
was been tested for 113 hours with average values of specific impulse and thrust of
1370 s and 400 plb respectively. [58]
Further work was performed on the tubular concept. A prototype which provided
547 pN of thrust at 75 %efficiency was constructed. The specific impulse, as measured
with time of flight, of the device was 1325 s. This device had extra electrodes for
electrostatic thrust vectoring. [26]
The United States Air Force launched a flight qualifying program for a complete
colloid thruster system. The design voltage was chosen to be substantially higher than
originally thought in order to take into account manufacturing variability, build up of
films on the emitters, the subsequent changes in the configuration of the propellant a
the tip and the effect of all this in the electric field. [27]Under the Colloid Advanced
Development Program, TRW designed a 432 needle colloid thruster for the US Air
Force. It had 12 modules each with 36 needles. Aiming for 1 mlb of thrust, it used
a solution of 3 g of Nal per 10 ml of glycerol. The motor operated only in positive
mode and the potential in the needles was 12.3 kV needles, while the extractor was at
-2 kV. It had a shield electrode around each needle to protect them from secondaries.
The specific impulse targeted was 1500 s. At TRW a, 36 needle module was tested
4350 hours and it showed a 15 % performance loss at end of life. The degradation was
caused by needle tar buildup, material build up on the shield electrode and needle
erosion. The three have a detrimental effect on the configuration and intensity of the
electric field. [59].
TRW went on to carry out investigations on multi needle configurations but needle
27
erosion or wax build ups were a recurring problem. The multi-needle arrangement
was complicated to reproduce while making sure that each needle had the right shape
for the field and also the same dynamic impedance for the flow.[9]
The following table 2.1.2 indicates some characteristics of the 432 needle prototype.
Table 2.1: Performance of a 432 needle prototype in 1974.
Thrust
1 mlb
Needle Voltage
12.2 kV
Specific Impulse
1365 s
Extractor Voltage
-2.3
Power
70 W
Specific Charge
12500 C/kg
Efficiency
58 %
Extractor Current
25 pA
There were concerns about the erosion and deposition on the needles.[75] From
1975 to 1978 work was done to understand these problems. An interim report already
pointed out that that the jet contained ions and that these ions could collide with
the drops and produce negative ions that were thus accelerate towards the emitter,
causing erosion. [51]
In 1978 an extensive report from the AFOSR was issued to try to identify the
mechanisms that caused the performance degradation of colloid thrusters that was
observed after 1000 hours of operation. The thrusters emitted only positively charged
drops and ions. The explanation put forward was that the fast ions hit drops and
produced some negative iodine ions that hit back the emitter. This, through erosion
and sputtering, changed the shape of the emitter, changing the electric field at the
tip of the emitters. The model seemed to account for the damage observed.
[43]
Erosion and deposition problems where technical issues that, while serious, could be
tackled by better designs. Issues of a different nature proved harder to overcome. The
following paragraph extracted from reference [75] sheds light on the type of concerns
of the program at the time:
"11.1 APPLICATION TO NEAR-TERM MISSIONS Hard targets for
advanced performance systems such as colloid propulsion are clouded by
28
two considerations that dominate current advanced space mission planning. First, is extreme dominance of budgetary considerations
and elimination of new "high risk" equipment wherever possible. Potential advanced capabilities are not used in the most preliminary mission
feasibility studies. This is especially true for housekeeping subsystems.
The second dominant considerationis the introduction of the Space Transportation System, or Space Shuttle. This has two effects. One is that the
cost of the Shuttle itself will dominate the available space budget for the
next several years. This is specially true for the NASA,
but also true
for DoD 's committed development of the Interim upper Sate. In an era
of fixed budgets for space, the payload funds will have to shrink to meet
any growth of Shuttle funding requirements. The payloads directly tied to
Shuttle, such as Spacelab and its derivatives, will likely fare better than
the free flyers. Second, the announced payload mass capabilities of
the shuttle are greater than current booster's. Many mission planners
assume that this extra capability will be used to extend the use of current
technology and relieve the dominant need for mass efficiency."
In other words, the colloid thruster was another casualty of Shuttle. Now it is
well known that Shuttle failed miserably to deliver on its promised economic efficiency
and that is another reason colloid thrusters, and electric propulsion in general, may
receive more attention.
2.1.3
Initial European electrospray-based colloid thruster development
The European Space Research Organization (ESRO), precursor to the current European Space Agency (ESA) started to work on colloid thrusters in 1969 and concentrated in developing a linear emitter of about 19 pin separation but first efforts
obtained still very low charge over mass (10 C/kg). [14]
A linear slit thruster using a pair of commercial razor blades was developed at
29
the University of Southampton. The device yielded a specific charge of 40000 C/kg,
a thrust of 100-150 pN at an efficiency of 60%. The performance was measured using
time of flight. The divergence of the jet was deemed too large. Another problem
identified was that the Nal glycerol propellant had undesirable electrochemical side
effects.
The liquid was fed into the slit by capillarity between the razor blades.
The two most typical gaps tested were 150 pm and 200 pm.
[41]
Due to the large
divergence of the beam, of about 600, a new design was made. It introduced a single
raised edge sandwiched inside a slit and made from a commercial razor blade. This
central blade was externally wetted and the emission took place at its rim. The
propellant wetted both sides of the blade and the beam divergence was reduced to
30'.[6] The thickness of the blade was 100 pm and the central edge protruded 400
pm from the edge of the exterior blades. The specific impulse was about 1530 s at an
operating voltage of 10.5 kV. The slit provided a thrust of 80 pN. The problem with
this design was the evaporation of propellant from the large exposed area.[42]
In Europe various slit colloid thrusters were tried both in linear and annular
configurations. A typical value of thrust per unit length was 70 pN/cm.[9] At the
European Space Research and Technology Center (ESTEC) an annular thruster was
tested and characterized. The propellant used was a mixture of sodium iodide and
glycerol of 20% by weight. Successful operation at both positive and negative polarity
was achieved. At high current levels in the negative mode hydrogen was liberated.
[5] In addition to the linear emitters, an annular emitter of 15 Pm annular gap was
designed. The voltages of operation where 12.5 kV and -2.6 kV. Using time of flight
a thrust of 0.5 mN at an specific impulse of 1223 s was measured. It was life tested
for 500 hours.[10]
As a result of an strategic study on the European propulsive needs, ESRO decided
that colloid thrusters was the most promising technology for secondary propulsion,
like station keeping. It chose to develop at a lesser level the field emission thruster
(FEEP) for missions requiring very high specific impulse. However there were concerns about the actual life colloid thrusters could to endure.[8]
Recognizing the interest of possible bipolar or AC operation, ESRO started a
30
program with the University of Southampton to find a propellant that would be
suitable for both polarities. A series of experiments using cyclic voltametry were
made to study the oxidation and reduction reactions that take place in the propellant
of choice at the time, NaI dissolved in glycerol. The results indicated that the said
propellant lacks an ability to accept electrons, making it unsuitable for continuous
negative mode operation. Then the efforts concentrated in trying to find a molecule
that would have the ability to take an electron or a hole, with equal ease. A candidate
material that is identified as a promising is 9:10 diphenyl-anthracene.[44].
Apart form the basic research on propellants, ESRO sponsored theoretical and
experimental studies of the energy losses from surface creation and viscous dissipation
of a single Taylor cone were performed.[2]
The University of Southampton tested a 3 cm linear slit thruster with a separation
between blades of 16 pm. A mixture of glycerol and Nal was used as propellant. The
current versus voltage plot exhibited quantum like jumps, probably due to bifurcations of the points of emission. Typical values of the operation of the were an specific
charge of 6000 C/Kg, a current of 100 pA, a thrust of 150 pN at an efficiency of 65%.
[3] This result, in 1975, was the last reported. It is unclear why no more publications
or reports on colloid propulsion where issued in Europe. A possible explanation could
be have to do with the reorganization of the Space supranational institutions in Europe that took place precisely in 1975. ESRO was incorporated in ESA and funding
and priorities shifted towards the development of the Ariane launcher.
2.1.4
Soviet electrospray-based colloid thruster development
In the Moscow Aviation Institute work on colloid thrusters went on for 35 years until
1995. Some of this work became public that year. A unit of 0.5-1mN thrust and
IONs total impulse consuming 30W was developed but, as far as it is known, never
flown. The thruster operated in DC and had an Isp of 104Ns/kg, an average specific
charge of 3330 C/kg, weighting 5kg and had a life of 2800 hours. The propellant was
glycerol-based. [63]
One of the applications that was looked at was small satellites. The propulsive
31
requirements of 35kg mass satellites ejected from the airlock of the the space station
MIR were studied. These satellites could be useful for conducting experiments that
require periods of astronaut interaction and also periods in a very clean acceleration
environment. The requirements were estimated to be a mass less than 7 kg, a total
power consumption of less than 30W, a thrust range of 4*10-4 to 10-3 N, and total
impulse 1400 Ns for an active flight time 30 days.
The study pointed to colloid
thrusters as the most efficient option for the mentioned requirements. [45]
2.1.5
Contemporary colloid thruster development
In 1997 in an article from JPL in which state of the art small propulsions systems
such as small ion engines and FEEPS was published. This work pointed out that:
of all micro-electric primary propulsion options
[...]
colloid thrusters are
quite possibly the most suited for microspacecraft primary propulsion applications[...]
[48].
Veterans of the technology of colloid thrusters, notably J. Perel, started a company named Phrasor in California. In 98 Phrasor proposed it again using an integrated MEMS design of "microvolcano emitter". They also proposed feeding the
emitters with capillarity. The foreseen application was microspacecraft main propulsion systems. [52]
At about that time a team formed by MIT, Yale and Busek, a company based in
Natick Massachusetts, published preliminary work on the topic.[46]
Development work was started in Stanford, where a prototype of 100 emitters
stainless steel capillaries with 0.006" od and 0.002" id. was constructed. [55]. The
prototype included the emitters, power processing unit and controls; it was tested for
100 hours. With a 4.4 kv acceleration potential it attained 200 s of Isp. The propellant
was glycerol with 2 molar sodium iodide. In addition to needle emitters, Stanford
also did experiments with a slit emitter provided by the company Phrasor.[56]
32
Applying modern results in electrospray applied to mass spectroscopy, electrospray
in the cone jet mode was proposed as a source of particles for colloid propulsion. This
allows for lower voltages of operation. An example: using a solution of Sodium Iodide
in formamide with a conductivity of 1.5 S/m. Experiments have been reported in
which at a massflow of 4 10-"
kg/s and an extraction voltage of 1547 V produces
an specific charge of 4760 C/kg. The thrust thus obtained is 0.134 /muN and the
specific impulse is 341 s.[18]
Busek worked on a 57 needle thruster with an extractor grid and accelerator grid.
A carbon nanotube neutralizer was included to emit electrons.
The needles where
made of stainless steel and had an internal diameter of 30 /mu m and an external
diameter of 180 /mum. The thrust was measured to be from 20 to 189 p N and the
maximum Isp was 400 s. The propellant was a Formamide solution with 0.5 S/m of
electrical conductivity. [25]
Interest in formation flying mission sand microspacecraft constellations revived
research on colloid thrusters. Missions like LISA require accuracies of the order of 0.1
p[N in a range up to 20 puN as well as high specific impulse (more than 500 s). Colloid
thrusters are specially well suited to meet the new requirements these new missions
have. According to an MIT study fine controllability and moderate to high delta V is
a niche that colloids could fill, having the additional advantage over other competing
technologies that they need substantially less power. [57]
With the aim of taking advantage of recent advances in miniaturization, MIT has
worked on a 1-D array of microfabricated emitters. This effort resulted in a prototype
using 240 emitters which, if operated at 4000V, would yield a performance 0.3 pN at
an Isp of 500 s, using formamide with Nal as propellant. The prototype was microfabricated on silicon.[71] Further work performed at MIT involved the microfabrication
of externally wetted 2D thrusters. Emission from the 2D emitter array was verified
experimentally. [70] [69] MIT has also advanced the understanding of what is called the
mixed ion-droplet regime, that is when the plume consists in a mixture of drops and
solvated ions. This is attractive because solvated ions have a charge/mass ratio too
large and drops only has too low Isp, only acceptable to low Delta V missions. Time
33
of flight and stopping potential experiments have been performed to determine the
energy distribution of Formamide Nal plumes.[36] Experimental results show that
drops have up to about 9000 C/kg. Solvated ions consist of Na+ ions and seven
molecules of formamide which results an specific charge of 280590 C/kg. While this
averaging effect can be useful, too much polidispersity in the beam can can result
in low overall efficiencies, [35] [37]. Following the design of Liquid Metal Ion Souces,
but using the ionic liquid EMI-BF4 instead of a liquid metal, externally wetted ion
sources have been developed at MIT[38]. The life of these devices has been shown
to increase dramatically if voltage is alternated instead of kept constant[39].
This
remarkable electrochemical behavior opens the prospect for bipolar thrusters based
in the technology of externally wetted ionic liquid ion sources.[40]
JPL has also worked to produce a 1 D array of 36 emitters on silicon. [49]
An article from the Edwards Air Force Base has proposed an AC colloid thruster
concept multiple acceleration gates using frequencies of around 10 MHz.[32]
A group from Tsinghua University in China has developed an integrated microthruster comprising a MEMS based micropump and a thruster made with print
circuit board technology (PCB) which has the advantage of being less costly than
MEMS. The propellant used was formamide with 30% Nal. An array of 81 emitters
of copper with an internal diameter of 0.3 mm and an external diameter of 0.5 mm.
The distance between emitters was 1.5 mm. Cantilever beam measurements of the
thrust indicated it was about 1.5 /mu N. The device was tested in air and suffered
from an unsteady burst-like mode of operation. [73] The design evolved into a MEMS
based using ICP etching. An array of 192 emitters was built using this technique.
An array of four emitters was tested on a vacuum chamber and measured 6.8 y N
with an extracting voltage of 2000 V in good agreement with theory. This thruster
is mentioned to meet an urgent requirement for a microsatellite. [77]
In the UK, interest in colloid thrusters has recently renewed. The main effort has
been the framework of a collaborative program between Queen Mary University of
London and the Rutherford Appleton Laboratory. [65] The focus of this work is to
produce emitters using microfabrication techniques. Single nozzles of 35 OD with 25
34
/Imu
m ID and 400 p, m length were manufactured. Using silicon microfrabrication
techniques which are highly scalable an array of 20000 emitters on a 75 mm 2 area has
been proposed. [67] The work is at the stage of testing individual emitters.[66].
35
36
Chapter 3
The needle-less thruster
An alternative design has been proposed to overcome two of the problems that colloid
thrusters had. The first problem is the alignment of the needles has to be very precise,
reproducible and durable [68] . The second and most important technical problem was
that, by design, the successful operation of the emitters relied on a sharp geometry
to operate. This geometry was often altered by erosion or deposition thus decreasing
the performance of the thruster.
[51]
We can decompose the functions of the needle in a conventional colloid thruster:
Fluidic To hold the fluid together.
Sharpness To provide a sharp equipotential tip that enhances the electric field.
Electrical To act as an electrode, that is to allow the transfer of electric charge to
the fluid.
The three functions are implemented by the same form, the needle or capillary.
Some benefit could come from decoupling these functions and assigning them to other
forms.
In terms of the electrical function it is clear that since the fluid itself is conductive,
the needle does not have to be. An electrode in contact with the fluid can provide
the necessary charge.
37
A measure of the sharpness of an edge or axially symmetric tip is its radius of
curvature R. Concerning the role of the sharpness of the needle it is worthwhile to
mention two things. First when the device is operating, the sharpness attained by
the Taylor cone is much more acute, a few nanometers, than any sharpening coming
from the geometry of the needle. Second, when no electric field is applied, for any
sharp internally wetted needle, the minimum radius of curvature is the radius of the
hemispherical liquid meniscus. Then a liquid cylinder of constant radius ending on a
hemispherical meniscus will be therefore more slender than any needle that surrounds
it. Any needle material around it does not improve the sharpness of the equipotential.
The ideal thing then would be to have the liquid held in place in vacuum in the shape
of a cylinder capped with a hemisphere. The third function of the needle is to keep
the fluid in place, it is desirable to hold it in a shape that is as slender as possible.
Ideally it would be held in place without any material around it, just in a vacuum.
This would be complicated to implement. The next best solution is to have a low
dielectric constant isolator solid holding the liquid in place. From an electrostatic
point of view its behavior is as close to vacuum as possible.
The author hypothesized that the function of holding the liquid in place may be
better accomplished by a hole on a low dielectric constant material.
This presents the additional attractive feature that alignment of the needles is no
longer an issue. Since arrays of holes can be made parallel to each other they remain
robustly aligned.
After looking at various materials that were insulators and had a low dielectric
constant, Teflon was selected because:
" It is electrically almost like vacuum, due to its low relative permittivity.
" It is extremely inert chemically.
" It has a very low surface energy, making it non-wetting.
Figure 3 shows a conceptual drawing of an array of holes in a block.
It is unclear to the moment why this constructive approach of using holes instead
of needles has not been used before. A valid concern that could be raised was that
38
Figure 3-1: Array of holes in a solid block
the startup voltage may be too high for the device to be attractive. In order to verify
if that was the case a series of computer simulations were run.
39
40
Chapter 4
Simulation
4.1
Statement of the problem
Of the many functions that needles have, as previously identified in chapter 3, enhancing the electric field at their tip is the one that is most often cited [4].
This
enhancement is supposedly needed to start up the instability at the tip of the liquid meniscus.
Therefore it was decided to perform a numerical simulation of the
electrostatic potential for two cases:
Case I: Sharp Needle A sharp stainless steel needle caped by a hemispherical perfectly conductive meniscus. The dimensions are shown in figure 4-1. The stainless steel is assumed a perfect conductor.
Case 1I: Hole on Teflon A hole of the same diameter on a Teflon, assumed a perfect insulator and of relative permittivity of 2.08.
The hole is capped by a
hemispherical perfectly conductive meniscus. The dimensions are shown in figure 4-2.
The geometry is axially symmetric. 3500 V are applied between the extractor and
the equipotential formed by: in Case I the needle and the hemispherical meniscus,
in Case II the liquid column and the hemispherical meniscus. We are interested in
comparing the intensity of the electric field at the tip for both cases. Note that due
41
2 mm
0.5
mm
1.55 mm
2.F5 mm
O.SO
mm
45
Figure 4-1: Case I dimensions
42
2'.T mm
I-1
Figure 4-2: Case II dimensions
43
to the linearity of the problem any value can be used for this voltage, as long as the
comparison between both cases is made with the same voltage.
4.2
Simulation method and accuracy
The electric potential and the modulus of the electric field were calculated using the
Student Version of the software program MaxwellTI. This program solves electrostatic problems using the finite element method coupled with an automatic adaptive
meshing algorithm that can almost arbitrarily increase the accuracy of the solution.
The target error was 1%. For Case I 0.55% was achieved after 8 passes, the
convergence data are summarized in figure 4-3. Case II, due to its much simpler
geometry, converged to an error of 0.667% at the first pass. Further refinement is
possible with this software, however it is unadvisable since, if forced to converge to
a smaller error, spurious results appear in the solution. This is due to the piecewise
linear approximation of the geometry of the hemispherical meniscus.
4.3
Simulation results
The main lesson that can be extracted from the results is that the electric field
enhancement takes place, not only in the case of the needle (Casel) but also in the
case of a hole in Teflon (Case II). In fact the modulus of the electric field at the tip
of the liquid meniscus is higher in Case II than in Case I, 3.97106 V/m compared to
3.53106 V/i.
Figures 4-4 and 4-5 show the electrostatic potential solution for cases I and II
respectively.
Figures 4-6 and 4-7 show the electric field modulus solution for cases I and II
respectively. The scale is the same for both figures and uses the maximum electric
field of case II as a reference. As we would expect, in both cases the electric field is
higher at the tip. What is more remarkable is that the value of the field near the tip
is slightly higher in Case II than in Case I.
44
Fercent &rrEnrrgy vs. Pss
12
10
6
4
2
a
-2
-4
I
3
5
4
6
Figure 4-3: Case I: Convergence
45
7
a
9
phi[V]
3.5000e+003
3.1500e+003
2.8000e+003
2.4500e+003
2.1000e+003
1.7500e+003
1.4000e+003
1.0500e+003
7.0000e+002
3.5000e+002
0.0000e+000
Figure 4-4: Case I: Electrostatic potential < for a metallic needle
46
phi(V]
=
F
3.5000e+003
3. 1500e+003
2.8000e+003
2.4500e+003
2.1000e+003
1.7500e+003
1.4000e+003
1.0500e+003
7.0000e+002
3.5000e+002
0.0000e+000
Figure 4-5: Case II: Electrostatic potential < for a hole in Teflon
47
E[V/m]:2
3.9688e+006
3. 5719e+006
3.1750e+006
"> 7782e+006
2.3813e+006
1. 9844e+006
1.5875e+006
1.1906e+006
7.93 7 6e+005
3.9688e+005
0.0000e+000
Figure 4-6: Electric field modulus for a needle, maximum electric field of 3.53106
V/m.
48
E(V/m]:4
3.9688e+006
3.5720e+006
3.1751e+006
2.7782e+006
2.3813e+006
1.9844e+006
1.5875e+006
1.1907e+006
7.9377e+005
3.9688e+005
2.7174e-009
Figure 4-7: Electric field modulus for a teflon hole, maximum electric field of 3.97106
V/m.
49
Figure 4-8: Model grid. For this particular run the dielectric material was teflon, the
hole had 0.4 m of radius and was placed 17.25 mm from the extractor plate.
4.4
Effect of the dielectric constant on the field
enhancement
The electric field at the tip of the meniscus is expected to decrease if materials with
higher relative permittivity than Teflon are used. In order to gain a quantitative
understanding of this phenomenon a series of models have been run. Figures 4-8 and
4-9 show a typical grid and close up of the electric field modulus near the tip.
50
Figure 4-9: Close up of the electric field modulus near the tip. For this particular run
the dielectric material was teflon, the hole had 0.4 m of radius and was placed 17.25
mm from the extractor plate.
Figure 4-10 shows the results of these simulations. For two values of the ratio
between the distance to the extractor, D, divided by the radius of the meniscus, R,
the inverse of the electric field at the tip normalized with the value for vacuum, has
been plotted as a function of the relative permittivity. The ordinate thus represents,
for a hole-based emitter, by how much the voltage has to be multiplied to obtain the
electric field at the tip that would be obtained if the propellant were held by itself in
vacuum. Figure 4-10 also illustrates the agreement with a previous analytical model
that approximated the geometry of the liquid column and meniscus by an sphere and
a cone[54]. It should be noted that the reduction in electric field enhancement at the
tip becomes more pronounced as the ratio D/R increases. This result can be useful
for the design of emitters from flat dielectric surfaces.
51
1.5-
*
D/R=6.88 (Maxwell Simulation)
1.45.
*
D/R=43.13 (Maxwell Simulation)
-
-D/R=43.13
(Theory from ref. 54)
1.4-
-D/R=6.88
I
(Theory from ref. 54)
1.35
N
0
.0
Cz
1.3W
E 1.250
1.21.15-
1.1 1.05-
1
3
5
7
9
11
Relative Permittivity
Figure 4-10: Effect of the relative permittivity on the Electric field at the tip.
52
Chapter 5
Experimental verification
An experiment was set up and operated by Dr. P. Lozano, Mr. P. Whitney and the
author to verify the feasibility of emission from a hole in Teflon.
The working fluid was Ethylene Glycol doped with LiCl to raise its conductivity
to 0.077 Si/m. The surface tension is 0.048 N/m and the dynamic viscosity 21 cP.
A digital syringe pump was used to provided a controlled volume flow to a tube
that fed a hole of 0.36 mm of diameter in a teflon block. In front of the hole at
a distance of 3 mm an extractor plate was placed. According to Equation 1.2 the
starting voltage would have to bee close to 2.92 kV, however that equation assumes
there is no Teflon around the liquid. An alternative way of deriving the startup
voltage is to calculate the electric field modulus required at the tip by equation 1.1.
The value of the electric field is 1.1 107 V/m. Then a Maxwell model of the geometry
can be run to identify the voltage that provides such an electric field at the tip. This
process yields a value of the starting voltage of 2.61 kV.
A range of voltages and mass flows was explored. A Taylor cone formed on the hole
and electrospray emission took place. Starting voltages were measured in the vicinity
of 3 kV. The most stable condition corresponded to a volume flow of 0.01pl/min and
4.36 kV applied on the liquid, keeping the extractor grounded.
Figure 5-1 shows a stable Taylor cone at the mentioned conditions.
The emission was remarkably stable.
When at some instances bubbles in the
stream disrupted the emission, it recovered by itself shortly.
53
Figure 5-1: Taylor cone formed on a hole on Teflon.
54
Chapter 6
Further work
Some potential issues that the needle colloid thruster could face, have been identified.
This chapter will present those and propose tentative solutions.
6.1
Exposure to the space environment
Teflon was chosen as a material among other things because of its good non wetting
surface properties. However, exposure to UV light or plasma can increase the surface
energy of teflon and make it sticky. If the design relies on the low surface energy of
the exposed teflon surface that could limit the life of the device.
Several studies have been made regarding the accelerated aging of Teflon due to
the exposure to the space environment. This interest stems from the fact that Teflon
is very often used as a coating in insulating materials for spacecraft. If thin layers of
Teflon are good to protect other materials it is reasonable to think that it can protect
itself to a certain extent.
The agents identified that contribute to the aging process are[17 [221:
" Thermal cycling
" Atomic oxygen
* Ultraviolet radiation
55
e Contamination
Thermal cycling and UV radiation would be less of an issue for a colloid thruster
because the extractor would shadow most of the Teflon.
Since the properties of
the propellants depend on the temperature, the thruster would probably have to
be thermally controlled and therefore thermal cycles would be mild. Therefore the
single most important factor that should be assessed experimentally is the effect of
atomic oxygen. Of special interest would be to know if the relative permittivity of
the surface is increased by the oxidation process. Also of concern is the possibility of
contamination of the surface by a conductive layer. If this layer became in contact
with the propellant it would adopt its potential and shield it from the electrode in
a very short time. A stationary simulation of this configuration including a 13 ptm
conductive layer yields, as expected, an electric field that is lower at the tip and
similar to the one in a parallel capacitor. Figure 6-1 shows a close up detail of the
field around the meniscus.
6.2
Hydraulic impedance
As discussed in section 1.3.2 the hydraulic impedance of the path between each emitter tip and the common feed manifold should be high enough. For a cylindrical hole
of diameter D and generatrix L, implies large L/D. This imposes a constructive constraint. Current laser drilling technology allows to drill arrays of holes of L/D of
about 120 on Teflon[33].
An alternative construction that would increase the flow impedance is proposed
and shown in figure 6-2. Instead of using one constant diameter for the tube, a flow
restriction is in series with the flow. The flow restriction may not have a large L/D
but it does increase the impedance. In the Figure 6-2 the restriction is placed at the
exit of the hole, but it could be placed anywhere in series. This construction would
be easier to manufacture on a single block by laser since the L/D is moderate for both
the channel and the flow restriction.
56
Figure 6-1: Close up of the electric field near the meniscus. The electric field intensity
that would be attained in the absence of the conductive film is 1.1 107 V/m.
57
Figure 6-2: Top: Cylindrical emitter hole. Bottom: Flow restriction at the outward
end of the cylindrical emitter hole to increase the flow impedance.
58
Chapter 7
Conclusions
The literature on colloid thruster development has been surveyed to identify the
technical reasons why the program was stopped in the 70s. From this bibliographical
research it has been understood that the performance was sensitive the sharpness
of the needles used as emitters. In normal operation of the thrusters, erosion and
deposition changed the shape of the needles and led to reduced performance.
An alternative way to construct an electrospray emitter for colloid thruster has
been introduced. Instead of using a sharp needle, a hole in a dielectric nonconductive
flat surface is used.
Numerical simulations have shown that the electric field can be higher using flat
holes than needles of the same inner diameter. Via numerical simulations the effect
of the relative permittivity of the dielectric material has been quantified.
Teflon, because of its low relative permittivity, has been identified as an apt material for such an emitter. Before holes on teflon can be applied for colloid thrusters
in space, further work will be needed to characterize the effects of potential deposits
and of the atomic oxygen on the electrical properties of Teflon. In any case, this type
of emitter can already be useful in some of the many applications where electrospray
is used.
59
60
Bibliography
[1] Research on the bipolar thrustor.
Air Force Aero Propulsion Laboratory-
Technical Report 110, 1967.
[2] A.G. Bailey. Natural limitations to the efficiency of colloid thrusters. 3rd European electric propulsion conference, Hinterzarten West Germany, 434, October
14-18 1974.
[3] A.G. Bailey.
Temperature effects and capillarity in an electrostatic liquid
thruster. AIAA Paper,434, 1975.
[4] A.G. Bailey. Electrostatic Spraying of Liquids. Research Studies Press LTD,
1988.
[5] A.G. Bailey, J.E. Bracher, and H.J. Robden. A capillary-fed annular colloid
thruster. J.Spacefraft, 9(7), 1972.
[6] A.G. Bailey, J.E. Bracher, and H.J. von Rohden. A comparative study of a
linear slit and a single-raised-edge colloid thruster. Electric propulsion of space
vehicles; Proceedings of the Conference, Abingdon, Berks., England, April 10-12
1973.
[7] J.C. Beynon, E. Cohen, D.S. Goldin, M.N. Huberman, P.W. Kidd, and S. Zafran.
Present status of colloid microthruster technology. AIAA Paper, 531, 1967.
[8] J.E. Bracher. Esro's activity on electric propulsion. Electric propulsion and its
space applications; Workshop, 2nd, Toulouse, France, A73-15712, 1972.
61
[9] A.W. Bright and B. Makin. The electrical propulsion of space vehicles. Contemporary Physics, 14(1):25-38, 1973.
[10] B.J.C. Burrows, R.G. Montague, T. Pedley, and M.F.A. Harrison. Endurance
testing of colloid thrustors. Electric propulsion of space vehicles; Proceedings of
the Conference, Abingdon, Berks., England, April 10-12 1973.
[11] W.C. Burson and P.C. Herren. Alternating current operation of a colloid source.
J.Spacefraft, 8(6), 1971.
[12] J. A. Carretero, M. Martinez-Sainchez, and F. Higuera. Numerical simulation of
a single-emitter taylor cone electrospray. Submitted to the J. of Fluid Mechanics,
2004.
[13] W.G. Courtney and C. Budnik.
Colloid propulsion using chemically formed
particles. AIAA PAPER, 254, 1966.
[14] T. Erin, B. Makin, D.J. Lines, and A.W. Bright. Fundamental phenomena of
the colloid thruster. European Space Research Organisation- ContractorReport,
22, 1970.
[15] J. Perel et al.
Electrodeless particle thrustor.
Air Force Aero Propulsion
Laboratory-Technical Report, 1967.
[16] C.F. Eyring, S.S. Makeown, and R.A. Millikan. Field's currents from points.
Physical Review, 31, 1928.
[17] M. Finckenor, J. Visentine, S. Adam, J. Zwiener, and V. Loebs. Contamination,
UV radiation, and atomic oxygen effects on ISS thermal control materials.
41st
AIAA Aerospace Sciences Meeting and Exhibit, Reno, NV, Jan, 2003-1084, 2003.
[18] M. Gamero-Castano and V. Hruby. Electrospray as a source of nanoparticles for
efficient colloid thrusters. Journal of Propulsion and Power, 17(5), 2001.
[19] J.W. Geis and J.M. Turner. Beam distribution effects on colloid engine performance. AIAA Paper, 1109, 1970.
62
[20] D.M.P. Gignoux. United states patent 3120736. Assignor to Cosmic Inc., February 1964.
[21] D.S. Goldin and G.L. Kvitek. An analysis of particle formation efficiency in a
colloid thrustor. AIAA PAPER, 253, 1962.
[22] K.K. De Groh, J.A. Dever, J.K. Sutter, J.R. Gaier, J.D. Gummow, D.A.
Scheimena, and C. He.
Thermal contributions to the degradation of Teflon
FEP on the Hubble Space Telescope.
46th International SAMPE Symposium
and Exhibition ; Long Beach, CA; USA; 6-10 May 2001., 2001.
[23] S.P. Harris. Research on charged colloidal particle propulsion. AIAA PAPER,
72, 1965.
[24] S.P. Harris and M. Farber. Development of a solid charged colloidal particle
thrustor. AIAA PAPER, 255, 1966.
[25] V. Hruby, M. Gamero-Castano, P. Falkos, and S. Shenoy. Micro newton colloid
thruster system development. 27th InternationalElectric Propulsion Conference,
Pasadena, CA, IEPC-01-281, October 2001.
[26] M.N. Huberman and S.G. Rosen. Advanced high-thrust colloid souces. J. Spacecraft, 11(7), 1974.
[27] F.A. Jackson. Progress in the development of a one-millipound-thrust colloid
propulsion system. Electric propulsion of space vehicles; Proceedings of the Conference, Abingdon, Berks., England, April 10-12 1973.
[28] J.P. Kesselring and H.S. Seifert. The use of preformed solid particles in colloidal
propulsion. AIAA PAPER, 727, 1967.
[29] P.W. Kidd. Parametric studies with a single-needle colloid thruster. J. Spacecraft,
5, 1868.
[30] P.W. Kidd. Parametric studies of a single needle colloid thruster. AIAA Paper,
530, 1967.
63
[31] P.W. Kidd, M.N. Huberman, and H. Shelton. A comparison of the time-of-flight
and thrust stand data for two 100 micropound colloid thrusters. AIAA Paper,
1114, 1970.
[32] D.E. Kirtley and J.M. Fife. A colloid engine accelator concept. AIAA PAPER,
3811, 2002.
[33]
Glen Arm MD Lenox Laser Company. Personal communication, 2003.
[34]
C.A. Low and W.R. Mickelsen. An electrostatic propulsion system with a direct
nuclear electrogenerator. Aerospace Eng., 21(58), 1962.
[35] P. Lozano. Studies on the Ion-Droplet Mixed Regime in Colloid Thrusters. PhD
thesis, Massachusetts Institute of Technology, 2002.
[36]
P. Lozano and M. Martinez-Sanchez.
Experimental study of colloid plumes.
AIAA PAPER, 3334, 2001.
[37]
P. Lozano and M. Martinez-Sanchez.
Experimental measurements of colloid
thruster plumes in the ion-droplet mixed regime. AIAA PAPER, 3814, 2002.
[38] P. Lozano and M. Martinez-Sanchez. Ionic liquid ion sources: Characterization
of externally wetted emitters. Submitted to the Journal of Colloid and Interface
Science, JCIS-04-498, 2004.
[39] P. Lozano and M. Martinez-Sinchez. Ionic liquid ion sources: Suppression of
electrochemical reactions using voltage alternation. Submitted to the Journal of
Colloid and Interface Science, JCIS-04-348, 2004.
[40] P. Lozano, M. Martinez-Sinchez, and J.M. L6pez-Urdiales. Externally wetted
ionic liquid thruster. 4th InternationalSpacecraft Propulsion Conference, Chia
Laguna, Italy, 2-9 June, 2004.
[41] M.R. Mahdavi, J. Lines, A.W.B. Bright, and B. Makin. Fundamental phenomena of the colloid thruster. European Space Research Organisation- Contractor
Report, 53, 1972.
64
[42] M.R. Mahdavi, B. Makin, and A.W. Bright. Optimisation of a single edge linear
colloid thruster. Electric propulsion of space vehicles; Proceedings of the Conference, Abingdon, Berks., England, April 10-12 1973.
[43] J.F. Mahoney, J. Perel, and B.E Kalensher. Final report-mechanisms of emitter
surface damage during electrohydrodynamic colloid particle generation and acceleration. Air Force Office of Scientific Research-TechnicalReport, 1026, 1978.
[44] B. Makin, D.J. Lines, and A.W. Bright. Electrochemical studies for selecting a
colloid thruster propellant. Electric propulsion and its space applications; Workshop, 2nd, Toulouse, France, June 21-23 1972.
[45] G. Malyshev, V. Kulkov, A. Shtyrlin, 1. Vishedkevich, and R. Bychkov. Comparative analysis of the propulsion system for the small satellites.
IEPC'95
InternationalElectric Propulsion Conference 24th, Moscow, Russia, 1995.
[46] M. Martinez-SAinchez, J. Fernandez de la Mora, V. Hruby, M. Gamero-Casta no,
and V. Khayms. Research on colloid thrusters. volume IEPC 99-014, October
1999.
[47] M. Martinez-Sainchez and J.E. Pollard. Spacecraft electric prolusion -an overview.
Journal of Propulsion and Power, 14(5), 1998.
[48] J. Mueller. Thruster options for microspacecraft: A review and evaluation of
existing hardware and emerging tehcnologies. AIAA PAPER, 3058, 1997.
[49] J. Mueller and C. Marrese et al. Jpl micro-thrust propulsion activities. AIAA
PAPER, 5714, 2002.
[50] J. Perel, T. Bates, J. Mahoney, and R.D. Moore. Research on a charged particle
bipolar thrustor. AIAA Paper,728, 1967.
[51] J. Perel and J.F. Mahoney. Interim report-mechanisms of emitter surface damage
during ehd colloid particle generation and acceleration. Air Force Aero Propulsion Laboratory-Technical Report, 0540, 1977.
65
[52] J. Perel and J.F. Mahoney.
Micro-electric propulsion using charged clusters.
Advanced Space Propulsion Workshop, 9th, Pasadena, CA, Mar. 11-13, 1998.
Proceedings, A98-31806, 1998.
[53] J. Perel, J.F. Mahoney, and A.Y. Yahiku. Analytical study of colloid annular
thrusters. AIAA Paper, 1113, 1970.
[54] P.Lozano, M. Martfnez-Sinchez, and J.M. L6pez-Urdiales. Electrospray emission
from nonwetting flat dielectric surfaces. Journal of Colloid and Interface Science,
in Press, 2004.
[55] F.M. Pranajaya and M. Capelli.
Progress on colloid micro-thruster research
and fligth testing. Annual AIAA/Utah State University Conference on Small
Satellites, 13th, Logan, UT, 3811, 2002.
[56] F.M. Pranajaya and M.A. Capelli. Development of a colloid microthruster for
flight demonstration on the emeral nanosatellite. AIAA PAPER, 3330, 2001.
[57] J.G. Reichbach, R.J. Sedwick, and M. Martinez-Sinchez. Micropropulsion system selection for precision formation flying satellites. AIAA PAPER, 3646, 2001.
[58] S.G. Rosen. Colloid and pulsed plasma thrusters for spacecraft propulsion. AIAA
Paper, 1254, 1973.
[59] S.G. Rosen. Life test/4350 hours/ of an advanced colloid thruster module. AIAA
Paper, 1078, 1973.
[60] R.D. Schutltz. United states patent 3157938. Assignor to Aerojet Corporation,
November 1964.
[61] R.D. Schutltz, L.K. Branson, and R. F. Cbaiken. United states patent 3122882,
March 1964. Assignors to Aerojet Corporation.
[62] H. Shelton and E. Cohen. Pulsed and alternatitng current colloid thruster studies. AIAA Paper, 178, 1970.
66
[63] A.F. Shtyrlin. State of the art and future prospects of colloidal electric thrusters.
IEPC'95 International Electric Propulsion Conference 24th, Moscow, Russia,
1995.
[64] S. Singer and M. Farber. Electro-propulsion with colloids. Astronautics, 1962.
[65] J. Stark, B. Kent, and B. Stevens et al. Colloid propulsion a re-evaluation, with
ain integrated design. AIAA PAPER, 4851, 2003.
[66] J. Stark, K. Smith, and S. Robertson. High accuracy measurements in electrospray source relevant to colloid thrusters. AIAA PAPER, 4847, 2003.
[67] J. Stark, B. Stevens, and M. Alexander. Fabrication and operation of microfabricated colloid thruster arrays. AIAA PAPER, 4852, 2003.
[68] K.W. Stark and A. Sherman. Present status of colloid microthruster technology.
NASA Technical Note, D-5305, 1968.
[69] L. Velisquez.
The design, fabrication and testing of microfabricated linear and
planarcolloid thruster arrays. PhD thesis, Massachusetts Institute of Technology,
2004.
[70] L. Velisquez, J. Carretero, A.I. Akinwande, and M. Martinez-S inchez. The
concept and development of a micro-fabricated colloid thruster array. volume
AIAA-2003-4850, July 2003.
[71] L.F. Velisquez-Garcia, A.I. Akinwande, and M. Martinez-Sainchez.
A micro-
fabricated colloid thruster array. AIAA PAPER, 3810, 2002.
[72] S.H. Wineland and R.E. Hunter. Negatively charged colloid generation research.
AIAA PAPER, 251, 1966.
[73] J. Xiong, Z. Zhou, X. Ye, X. Wang, Y. Feng, and Y. Li. A colloid micro-thruster
system. Micro-electronic Engineering, 61-62, 2002.
[74] A.Y. Yahiku, J.F. Mahoney, and H.L. Daley. Experimental study of colloid
annular thrusters. AIAA Paper, 1112, 1970.
67
[75] S. Zafran. Colloid advanced development program. NASA Tech. Report, 14009,
1977.
[76] S. Zafran and R.F. Kemp. Colloid microthruster test stand. AIAA Paper,314,
1969.
[77] Zhaoying-Zhou, Jijun-Xiong, and Xiongying-Ye et al. Designing, fabrication, and
test of a mems colloid thruster. SPIE-Int. Soc. Opt. Eng, 4928: 148-54, 2002.
68